Samsung and SK Hynix Lead South Korea's $518B Chip Plan

An $880bn industrial bet

 

On 29 June 2026, President Lee Jae Myung appeared alongside the chairmen of Samsung Electronics and SK Hynix to announce one of the largest industrial commitments of the AI era. Reuters and CNN reported the government describing the plan as a "great leap forward," built on what it called a "triple axis" of semiconductors, physical AI and data centres. The chip investment alone was put at more than $576 billion over several years.

The commitment breaks down into three parts:

• Chip fabs: Samsung and SK Hynix will spend a combined 800 trillion won, about $518 billion, to build two new fabrication plants each in the country's southwest. Samsung has chosen the city of Gwangju for one of its sites.
• AI data centres: the science ministry set out plans for 550 trillion won of investment by 2029, rising above 1,000 trillion won by 2035.
• Combined figure: Bloomberg put the chips-and-data-centres total at roughly 1,350 trillion won, or about $880 billion.

The size of the package matters. The more important point is what the plan tries to connect. It ties chip manufacturing to the development of a region outside the crowded capital area, and both to the promise of cleaner power for an industry under growing pressure to decarbonise. Those goals usually sit in different ministries and different corporate strategies. Putting them in one blueprint is the ambitious part, and the part that will be hardest to deliver.

 

Why power and water now shape chip strategy

 

The demand behind the investment is not speculative. The International Energy Agency expects global data-centre electricity use to more than double to around 945 terawatt-hours by 2030, slightly more than Japan consumes today, with demand from AI-focused facilities rising fastest. Samsung and SK Hynix are central to meeting it. Together they make about two-thirds of the world's memory chips, and SK Hynix is a leading supplier of the high-bandwidth memory that advanced AI systems rely on.

Making those chips is resource-heavy, which is why an expansion on this scale is now as much an energy question as a manufacturing one. A single advanced fab can consume around 10 million gallons of ultrapure water a day, comparable to a small city, according to figures cited by the World Economic Forum. Asset manager Robeco points to research finding that more than 40% of fabs announced since 2021 sit in areas expected to face high or extremely high water stress by 2030. Energy demand is on a similar path: one widely cited estimate put the semiconductor industry's electricity use at about 149 billion kWh in 2021.

The relevance for sustainability leaders outside Korea is direct. Samsung and SK Hynix supply most of the world's major AI and electronics value chains, so however their new capacity is powered, the emissions will surface in other companies' supply chains, far from Gwangju.

 

What the state is putting on the table

 

Most of the coverage went to what the companies are spending. The government's contribution is arguably more distinctive, and it is not money.

Seoul is offering speed. The trade ministry has said it will compress the process from licensing to construction, treating slow permitting as a competitive weakness to fix. The plants are also being steered toward the southwest, away from the semiconductor belt around Seoul that is already short of land and power. That location has drawn domestic political debate, with opposition politicians questioning whether it was chosen on industrial grounds or for regional politics, while the government casts it as central to its push to rebalance growth beyond Seoul. Beyond the fabs themselves, the plan takes in a robotics and components cluster on the west coast, meant to give Korea a foothold in physical AI.

The companies bring capital and manufacturing depth that few rivals can match, along with climate commitments they made well before this announcement. Both belong to RE100, the global campaign for fully renewable electricity, and both have set net-zero targets. Neither chairman played down the difficulty. SK Hynix's Chey Tae-won said the project would need large sites and enough power, water and skilled workers, and pointed out that the company's existing cluster in Gyeonggi Province took nine years to build. Building the resource questions into the plan from the start, rather than leaving them for later, is a real departure from how chip expansions have often worked.

Officials have gone further, suggesting the southwest's renewable resources could give the chipmakers an advantage as customers and regulators demand cleaner power. That is where the growth story and the climate story are meant to converge. The evidence for it is mixed.

 

The renewable-energy gap

 

South Korea starts from an unusual position for a technology leader. Analysis from the Institute for Energy Economics and Financial Analysis put renewables at roughly 9.6% of the country's power mix in 2023. The Climate Action Tracker describes South Korea as having one of the lowest shares of renewable generation in the OECD and G20. That is the backdrop against which a $500 billion clean-industrial ambition has to be judged.

The decision to build in the southwest does have logic behind it. The Honam region that takes in the southwest is among the country's richest in renewable resources. Government-linked and independent assessments put around 40% of national solar capacity there, and Jeolla province accounts for roughly a third of the country's renewable generation. Locating power-hungry fabs nearer to where clean electricity is strongest is a reasonable starting idea.

The corporate targets are real, and moving in the right direction from a low base. Samsung joined RE100 in 2022 and has committed to net-zero emissions across its own operations by 2050. Its renewable-energy use reached 10,069 GWh in 2024, more than double the 2020 figure, and it reported a renewable share of about 31% in 2022. SK Hynix was the first memory chipmaker to join RE100, in November 2020, and aims for net zero and fully renewable power by 2050, with an interim goal of 33% renewable electricity by 2030. Its own disclosures show its renewable share climbing from 4% in 2021 to about 30% a year later. Both companies have also invested in water recycling and in partnerships with local governments to reuse treated municipal water at existing plants.

Both firms have made real progress quickly, and both have said plainly that Korea is one of the harder places in the world to buy renewable electricity. Closing the gap between a roughly one-third renewable share today and a fully renewable target is exactly what this partnership is meant to achieve.

 

What still needs to be proven

 

It is too early for verdicts. The questions worth tracking are about execution.

The first is the grid. Renewable-rich regions, in Korea and elsewhere, have gone through spells of curtailment when generation outruns the transmission network's capacity to move it. How quickly new transmission and interconnection capacity gets built will determine how much of the southwest's clean power the fabs can actually use. Tied to this is the problem of round-the-clock supply. Fabs never stop; solar and wind do. So the clean-power outcome depends on the mix eventually chosen between renewables backed by storage and firmer sources such as nuclear, and that mix is not yet decided.

Water is a separate constraint. The southwest's supply is smaller than that of Korea's major river basins, and much of it is already spoken for. Whether recycling and municipal partnerships can meet demand at fab scale is an open question of engineering and governance.

There is also a wider one. If Korea can align national development goals, private capital and corporate climate targets inside a single programme, other governments under the same AI-infrastructure pressure will want to know how it was done. If the pieces end up moving at different speeds, that will tell them something too.

None of this makes the plan a clean win for growth or for the climate, and it was never likely to be. What South Korea has committed to is really a race between two build-outs. Private capital can raise fabs and data centres fast. The harder job belongs to the state and the utilities: laying down the grid, the water systems and the clean-power supply quickly enough that all this new capacity does not end up locking in a larger fossil footprint by default. The investment figures are now settled. Whether clean power arrives in time is not, and that is the part worth watching.

 

Sources: CNN, CNBC, Bloomberg, Samsung, SK Hynix, Climate Action Tracker, WEF, Robeco, IEEFA